21 Comments

John,

Thanks for the great job you did for the whole of the summer meet. You have missed your true calling, that of herding cats.

K.J.

Expand full comment
author
Aug 6, 2023·edited Aug 6, 2023Author

TY. That point has been brought up before — probably as there is some truth to it.

Expand full comment

Did you consider asking you’re talking to a bot? I guess it would take a bit of nerve. If it was a person talking from a script then that person might repeat the script or be thrown off script.

Expand full comment
author
Aug 3, 2023·edited Aug 4, 2023Author

Bill: I did not, as I wouldn't expect a straightforward answer. It's an interesting thought though.

Expand full comment

Two comments:

1. Yeah - we already know that Chat GPT lies.

2. If you tell it so early in the conversation that you know it's a bot - you teach it to correct itself to more effectively deceive you.

Expand full comment
author

Good point in #2 — which means it would be best NOT to tell it that you know it is AI...

Expand full comment

This reminds me of all the fake accounts Elon apparently cleaned up out of twitter. Those accounts, or "bots" were responsible for driving a large portion of comments and retweets. I also remember Rush Limbaugh who hired a team to investigate the origins of disparaging online social media posts and they traced it back to a few people working in their parents basements!!! Now they've replaced actual people with AI. Don't get me started not the deep fake videos....we'll never be able to trust a video anymore. See this widely popular Tom Cruise deep fake video: https://www.youtube.com/watch?v=lhi31-pOz5M. Here's the TikTok website with more: https://www.tiktok.com/@deeptomcruise?lang=en

Expand full comment
author

Jo: Thank you for your usual good insights...

Expand full comment

John,

Glad to see flattery will get one nowhere with you.

K.J.

Expand full comment
author

That depends. If it was a REAL hottie, maybe it would!

Expand full comment

Every time an exceptionally good woman contacts me...

...it’s always a scam

Expand full comment

Sorry, this is non-responsive to your actual article, but I am here and I have a moment, so I thought I would go ahead and write this. I dabbled in ChatGPT for the first time a couple of days ago. I asked it a series of questions on a subject about which I am very very familiar. The first answer was incomplete so I provided a follow up question. The answer to the follow up question was flat out wrong so I moved on and asked a completely different follow up question. The answer to that question was utterly and completely non-responsive to that question and instead the AI said that it had thought some more about my first follow up question and wanted to make some corrections. Interesting. Why didn’t it answer correctly the first time? It is almost as if it started to figure out that I knew what I was talking about and I needed to set the record straight straight to develop some degree of credibility. But in my mind, the credibility is completely gone.

Expand full comment
author

Nina: I hope you read my four-part series on AI, where I went into great detail. We are on the edge of some very sophisticated AI, which will open the door to some very bad behavior.

Expand full comment

Artificially determines aspects of the targeted individual during initial interactions to determine the degree of propaganda it is able to inflict without awareness...

...prior to that the AI can only evaluate static data vs. the sequences and complexities of the questions or tasks asked...

Expand full comment

Thank you for replying, Pete. My limited experience confirms what you just wrote. How insidious. AI is not meant to help us. It is meant to con us. Like the drivel propaganda that Wikipedia morphed into after it was assimilated. Yes, I know, some say that was the plan with Wikipedia all along. I tend to think Wikipedia started out well meaning but then got infiltrated, but I could be wrong. It could have been evil from the start, just like AI.

I am an attorney. By now every attorney on the planet, excepting only those who are living under rocks, know that AI can't be trusted. Still I thought I would try for myself rather than take someone else's word for it. Yep, I learned for myself that AI can't be trusted.

Expand full comment

"...If you're interested in what the CIA wants you to think about a subject, check it out on Wikipedia. The Agency has been covertly "moderating" it for over ten years..." - https://markcrispinmiller.substack.com/p/if-youre-interested-in-what-the-cia

Expand full comment

As an Engineer working on energy security for the military I have had two very similar experiences with two attractive oriental females. Having spent a fair time in Asia with similar young ladies, it was seductive initially. By asking questions and carrying the dialogue two steps further, the same reality of what was probably happening dawned on me and I broke it off.

Expand full comment

Very good analysis. Since I first learned to.program in 1961 I have never trusted a computer output where I couldn't compile with a trace and verify every mathematical step.

R. Motz

Expand full comment

This is rich! It also proves your points about AI. Sadly, AI will be able to fool msot of the people most of the time.

Expand full comment
Comment deleted
Expand full comment
author

Yes, This is a progression on prior shenanigians. Yes Linkedin (which prides itself on screening out mis-information) should screen out blatant deception.

Expand full comment